renewable energy integration: technological and market design
TRANSCRIPT
Renewable Energy Integration: Technological and Market Design
Challenges
PSERC Future Grid Webinar February 19, 2013
Shmuel Oren, Duncan Callaway, Anthony Papavasiliou, Johanna Mathieu
(UC Berkeley) Timothy Mount, Robert Thomas, Max Zhang
(Cornell University) Alejandro Dominguez-Garcia, George Gross (University of Illinois at Urbana/Champaign)
A PSERC Future Grid Initiative Progress Report
PSERC Future Grid Initiative • DOE-funded project entitled "The Future Grid to Enable Sustainable
Energy Systems”(see http://www.pserc.org/research/FutureGrid.aspx)
• Focus of this webinar: Accomplishments in the thrust area “Renewable Energy Integration: Technological and Market Design Challenges”
• Objectives of this thrust area: Explore technological solutions, market design, resource dispatch tools and new planning frameworks for dealing with the uncertainty and variability of intermittent renewable generation resources. Task objectives include: • Develop distributed control paradigms and business models for mobilizing
demand response to mitigate the uncertainty and variability introduced by massive integration of renewable energy resources
• Develop market mechanisms that will incentivize load response and flexibility and correctly price uncertainty (or uncertainty reduction) on the demand and supply side.
• Develop dispatch and planning tools that can explicitly account for uncertainty, variability and flexibility (e.g. storage) in resource optimization and reserves procurement.
• Develop simulation tools that can account for increased uncertainty in verifying system and market performance 2
Context and Motivation
3
Uncertainty
4
Negative Correlation with Load
0
50
100
150
200
250
win
d po
wer
out
put
(MW
)
24 48 72 96 120 144 168 3000
4000
5000
6000
7000
8000
load
(MW
)
hour
wind power
load
5
All Rights Reserved to Shmuel Oren
Conventional Solution
Source: CAISO 7
I Need a Brain
Source: GE 8
• Accounting explicitly for uncertainty in operation and planning • Stochastic unit commitment (with endogenous reserves
determination) to support renewable penetration and demand response
• Probabilistic planning and simulation models (accounting for renewables, storage and demand response
• Mobilizing demand response (DR) and a paradigm shift to “load following available supply” provides an economically viable and sustainable path to a renewable low carbon future. • Price responsive load • Energy efficiency • Deferrable loads:
• EV/PHEV • HVAC • Water heaters • Electric space heaters • Refrigeration • Agricultural pumping
9
Thermostatically Controlled Loads (TCLs)
Making the Grid Smarter
10
Planning and Market Design for Using Dispatchable Loads to Meet Renewable Portfolio
Standards and Emissions Reduction Targets
Direct and Telemetric Coupling of Renewable Energy Resources
with Flexible Loads
Mitigating Renewables Intermittency Through Non-
Disruptive Load Control
Max Zhang, Tim Mount and Bob Thomas Cornell University
George Gross and Alejandro Dominguez-Garcia
with Yannick Degeilh University of Illinois at Urbana-
Champaign
Shmuel Oren with Anthony Papavasiliou
UC Berkeley
Duncan Callaway with Johanna Mathieu and Mark Dyson
UC Berkeley
Probabilistic Simulation Methodology for Evaluating the
Impact of Renewables Intermittency on Operation
and Planning
Today’s Presentations on Future Grid Tasks
Task 1: Direct and Telemetric Coupling of Renewable Energy Resources with Flexible Loads
Shmuel Oren Anthony Papavasiliou
UC Berkeley
11
Alternative Demand Response Paradigms
12
13
Alternative Demand Response Paradigms
Evaluation Methodology
• Comparison of alternative approaches to flexible loads mobilization requires explicit consideration of uncertainty for consistent determination of locational reserves.
• Stochastic unit commitment optimization accounts for uncertainty by considering a limited sample of probabilistic wind and contingency scenarios, committing slow reserves early with fast reserves and demand response adjusted after uncertainties are revealed.
• Economic and reliability outcomes are calculated using Monte Carlo simulation with large number of probabilistic scenarios and contingencies
14
Results and Benefits
• Developed a stochastic optimization method for efficient reserves deployment in an environment with high renewables penetration
• Developed a consistent method for assessing alternative demand response integration strategies (direct coupling vs. market)
• Advanced the state of the art for stochastic unit commitment at practical scale employing High Performance Computing (HPC)
• Study of how computational time for stochastic unit commitment scales with number of processors and solution accuracy
15
Model Structure
16
Unit Commitment
17
Two Stage Stochastic Unit Commitment
18
Scenario Sample Selection
19
New Scenario Selection Method
20
Wind Modeling and Data Sources
21
Load Variation Represented by Day Types
22
Parallelization and HPC Application
23
California Case Study
• Stochastic Optimization captures nearly 50% of gains under perfect forecasting of load and wind outcomes
• Direct coupling marginally more expensive than a centralized market but reduces load shedding due to better representation of load flexibility
• Transmission constraints can play a significant role in determining cost and resource adequacy
24
Computational Efficiency Study
25
References • Papavasiliou Anthony, Shmuel Oren and Richard O’Neill, “Reserve Requirements for Wind Power
Integration: A Scenario-Based Stochastic Programming Framework”, IEEE Transactions on Power System, Vol 26, No4 (2011), pp. 2197-2206
• Papavasiliou A. and S. S. Oren, ”Integrating Renewable Energy Contracts and Wholesale Dynamic Pricing to Serve Aggregate Flexible Loads” Invited Panel Paper, Proceeding of the IEEE PES GM, Detroit, Michigan, July 24-28, 2011.
• Papavasiliou A. and S. S. Oren “Integration of Contracted Renewable Energy and Spot Market Supply to Serve Flexible Loads”, Proceedings of the 18th World Congress of the International Federation of Automatic Control, August 28 – September 2, 2011, Milano, Italy.
• Papavasiliou A.and S. S. Oren, “Stochastic Modeling of Multi-area Wind Power Production “, Proceedings of PMAPS 2012, Istanbul Turkey, June 10-14, 2012.
• Oren S. S., Invited Panel Paper ” Renewable Energy Integration and the Impact of Carbon Regulation on the Electric Grid “, Proceeding of the IEEE PES GM, San Diego CA, July 22-26, 2012.
• Papavasiliou A., S. S. Oren, ” A Stochastic Unit Commitment Model for Integrating Renewable Supply and Demand Response” Invited Panel Paper, Proceeding of the IEEE PES GM, San Diego, CA, July 24-28, 2012.
• Papavasiliou A., S. S. Oren, “Large-Scale Integration of Deferrable Demand and Renewable Energy Sources in Power Systems”, Accepted for publication in a special issue of the IEEE PES Transaction.
• Papavasiliou A., S. S. Oren, “Multi-Area Stochastic Unit Commitment for High Wind Penetration in a Transmission Constrained Network”, Accepted for publication in Journal of Operations Research.
• Papavasiliou Anthony, Shmuel Oren, Barry Rountree “Applying High Performance Computing to Multi-Area Stochastic Unit Commitment for Renewable Energy Integration”, Submitted to Mathematical Programming (February 2013)
26
Mitigating Renewables Intermittency Through Nondisruptive Load Control
Duncan Callaway Johanna Mathieu
Mark Dyson University of California, Berkeley
Notes: Callaway is the Task leader, [email protected]. Mathieu is currently a postdoctoral scholar in the Power Systems Laboratory at ETH Zurich. Dyson was not
funded by the project, but contributed to the resource assessment.
27
Context
• Renewables integration requires power system flexibility (e.g., managing frequency response and energy imbalances)
• Centralized control of load resources could be a low cost solution: the grid connected resources exist already
• But the costs could be pushed upward by: • Communications & metering infrastructure
requirements (system operators need high quality telemetry data in certain applications)
• Customer payments (if end-use function has to be seriously compromised)
28
Research Goals
• New methods to model and control aggregations of thermostatically-controlled loads (TCLs) that • Reduce communications and
power measurement requirements • Minimize temperature deviations
• Evaluate how different real time communications abilities affect • Ability to accurately estimate local temperature and
ON/OFF state of loads • Controllability of load resources
• Analyze TCL resource potential, costs, and revenue potential associated with TCL control
TCLs
29
Basic Residential TCL Control Architecture
30
• All control occurs within existing TCL temperature deadband
• Use substation SCADA to measure aggregate power consumption
• Estimate states in aggregation model
• Broadcast control signal, possibly via AMI
dispatch instructions
• Loads receive broadcasted control signal and, based on current load temperature, turn ON, OFF or remain in current state
30
Aggregated TCL Model ‘State bin transition model’
…
…
ON
OFF
normalized temperature
…
stat
e
[Similar to that proposed by Lu and Chassin 2004; Lu et al. 2005; Bashash and Fathy 2011; Kundu et al. 2011]
31
…
…
Consider thousands of TCLs traveling around a normalized temperature dead-band.
ON
OFF
normalized temperature
…
stat
e
Aggregated TCL Model ‘State bin transition model’
32
Divide it into discrete temperature intervals.
ON
OFF
normalized temperature
…
…
stat
e
Aggregated TCL Model ‘State bin transition model’
33
ON
OFF
normalized temperature
…
…
stat
e
Aggregated TCL Model ‘State bin transition model’
Forcing the system: decreasing aggregate power. 34
ON
OFF
normalized temperature
…
…
stat
e
Aggregated TCL Model ‘State bin transition model’
Forcing the system: increasing aggregate power. 35
Question: How important is real time metering?
36
• Reference case: Meter power and temperature at all controlled loads, error following dispatch signal = 0.6% RMS (smaller is better)
• Case 1: Meter the ON/OFF state at all loads, measure aggregate power at the distribution substation. Result: error = 0.76% RMS
• Case 2: Meter only aggregate power at distribution substation. Result: error = 5% RMS • Note, this error compares favorably to conventional generators
All results assume: • 17 MVA substation load • 15% of load (1,000 TCLs) is controlled • Aggregate power measurements include all loads on substation • Total substation load can be forecasted with 5% average error
on a one minute horizon
Answer: Not important; state estimation works
How LARGE is the Resource Potential? Estimates for most of California (5 largest utilities) based on Renewable
Energy Certificates and California Energy Commission data.
2012 Resource Duration Curve
2020 Resource Duration Curve,
assuming increased efficiency and 30% of water/space heaters converted to electric
37
2000 4000 6000 80000
2
4
6
8
10
12
14Energy Capacity
GW
h
Hours2000 4000 6000 8000
0
10
20
30
40
50Power Capacity
GW
Hours
Central ACHeat pumpsWater heatersRefrigerators
2000 4000 6000 80000
2
4
6
8
10
12
14Energy Capacity
GW
h
Hours2000 4000 6000 8000
0
10
20
30
40
50Power Capacity
GW
Hours
2000 4000 6000 80000
2
4
6
8
10
12
14Energy Capacity
GW
h
Hours2000 4000 6000 8000
0
10
20
30
40
50Power Capacity
GW
Hours
A. 2012 Base Scenario
B. 2020 Base Scenario
C. 2020 Electrification Scenario
2000 4000 6000 80000
2
4
6
8
10
12
14Energy Capacity
GW
h
Hours2000 4000 6000 8000
0
10
20
30
40
50Power Capacity
GW
Hours
Central ACHeat pumpsWater heatersRefrigerators
2000 4000 6000 80000
2
4
6
8
10
12
14Energy Capacity
GW
h
Hours2000 4000 6000 8000
0
10
20
30
40
50Power Capacity
GW
Hours
2000 4000 6000 80000
2
4
6
8
10
12
14Energy Capacity
GW
h
Hours2000 4000 6000 8000
0
10
20
30
40
50Power Capacity
GW
Hours
A. 2012 Base Scenario
B. 2020 Base Scenario
C. 2020 Electrification Scenario
Potential Revenues for Regulation and Load Following (per TCL per year)
*Results depend on the climate zone
Note: cost requires a separate analysis!
Air conditioners* Heat pumps* Combined AC/HP* Water heaters Refrigerators
Regulation
$9-79 $100-170 $160-220
$61 $25
Load Following
$2-9 $9-14
$16-18 $35 $14
38
Uses and Potential Benefits of Results
• Reduced cost to deploy centralized control of loads on distribution circuits • AMI could broadcast control signals • Substation SCADA may be all that is required for real
time measurement • Roadmap for which loads are best for fast
demand response • Electrification of heating has big benefits
• Results lay groundwork for demonstration • Currently in discussion with several load aggregators
to run a pilot
39
References
• Mathieu, J.L.; M.E. Dyson, and D.S. Callaway. Using Residential Electric Loads for Fast Demand Response: The Potential Resource, the Costs, and Policy Recommendations. To appear in the Proceedings of the 2012 ACEEE Summer Study on Energy Efficiency in Buildings, Pacific Grove, CA, August 12-17, 2012.
• Mathieu, J.L.; S. Koch, and D.S. Callaway. State Estimation and Control of Electric Loads to Manage Real-Time Energy Imbalance. IEEE Transactions on Power Systems (in press), 2012.
• Mathieu, J.L.; and D.S. Callaway. State Estimation and Control of Heterogeneous Thermostatically Controlled Loads for Load Following. Proceedings of the 45th Hawaii International Conference on System Sciences (HICSS45), Wailea, Hawaii, January 4-7, 2012.
40
Cornell Research Team: K. Max Zhang
Timothy D. Mount Robert J. Thomas
41
Summary
42
Premises: 1) Need improved methods for evaluating the operations
and planning of a future grid with: a. High penetrations of renewable sources b. Storage and deferrable demand*
2) Establishing public support for the future grid will require that customers see tangible benefits (e.g., lower bills)
* Decouples the purchase of electricity from the delivery of an energy service. Contributions: 1) Developed an integrated multi-scale physical and
economic framework for modeling deferrable demand 2) Evaluated the effects of stochastic renewable sources
and deferrable demand on total system costs and emissions from generating units (to be completed)
Accomplishments 1) Developed a model for aggregating customers with electric
vehicles to determine the maximum hourly charge limited by the capacity of chargers and commuting patterns
2) Developed a model for aggregating buildings with thermal storage capabilities for cooling limited by the capacities of compressors and the energy stored
3) Used #1 and #2, in addition to stochastic wind generation, as inputs into the SuperOPF, a stochastic form of multi-period Security Constrained Optimal Power Flow (SCOPF)
4) Augmented the SuperOPF by adding emission/damage coefficients to the operating costs of generating units
5) Used #1- 4 to simulate the system effects and costs of using deferrable demand to mitigate the variability of generation from wind sources and reduce emissions (to be completed)
43
Context of the Research:
An Integrated Multi-Scale Framework
44
SuperOPF Costs
PEV charger capacities Commuting Patterns Nodal Capabilities
Ice storage systems Buildings Nodal Capabilities
Stochastic wind at 16 sites
North East Test Network
70/30 Level I/II 50/50 Level I/II 30/70 Level I/II
• A 20% penetration of commuters are assumed to use PEVs in the NPCC region. • 80% of the PEV load is assigned to valley hours to take advantage of the low prices. • The remaining 20% is assigned to shoulder and peak hours to reduce ramping costs • The Charging Flexibility Constraint (CFC) may restrict PEV charging during the morning
commuting hours Valentine, Temple and Zhang (2011) J. Power Sources
Results I: System Load with “Intelligent” Charging of PEVs under Different
Penetrations of Low and High PEV Charging Rates
45
Results II: Percent Reduction in Wholesale Energy Costs Using “Intelligent”
Charging and Valley-Fill Charging vs. Charging-at-Will
Intelligent Charging
Valley-fill Charging
• Charging-at-Will (i.e., commuters charge their PEVs as soon as they arrive at home) is the baseline. • Intelligent charging results in significant reductions in the costs compared to charging-at-will and
valley-fill charging. • Percentage reductions in the costs with intelligent charging are higher with higher penetrations of
PEVs because more PEVs provide a greater capability for modifying wholesale prices and the system cost of ramping.
Valentine, Temple and Zhang (2011) J. Power Sources 46
Results III: System Load and System Cost Reductions with Price-Responsive Ice Storage Systems
• The system benefits of aggregating Ice Storage Systems in large commercial and industrial buildings in New York State were evaluated for the NYCA.
• Heuristic methods were used to reduce system costs in a two-settlement wholesale market that accounts for both the steady-state costs and ramping costs of generating units.
• The optimal management of thermal storage significantly reduces both the peak load and total system costs, and also flattens out the daily load profile of conventional generating units. Palacio et al. In preparation
47
Results IV: Structure of the SuperOPF
OBJECTIVE FUNCTION FOR DAILY OPERATIONS: Minimize the expected cost of meeting load, including the costs of load-not-served, reserves and ramping, for a set of credible system states over 24 hours subject to: 1) Network constraints, 2) Stochastic wind inputs, 3) Contingencies. DETERMINE THE OPTIMUM NODAL VALUES FOR: 1) Hourly dispatch profiles for generating units, 2) Amount of wind dispatched/spilled, 3) Amount of Load-Not-Served, 4) Up and down reserve capacity, 5) Amount of ramping, 6) Nodal energy prices, 7) Charging/discharging of deferrable demand, 8) Charging/discharging of collocated storage,
48
APPLIED TO: North East Test Network (NET Net)
Hot Summer Day
Results V: Optimum Daily E[Pattern of Generation] for the NET Net
with and without Wind Capacity Installed
49
Case 1: Base Case 2: Base + 29GW Wind
Case 1: Ramping for the daily load profile is provided by oil and natural gas capacity. Case 2: Wind displaces mainly oil and natural gas capacity, but this conventional capacity provides the additional ramping services needed to mitigate the uncertainty of wind generation. Some wind is spilled.
Results VI: Optimum Daily E[Pattern of Generation] for the NET Net
with Deferrable Demand or Collocated Storage
50
Case 3: Base + 29GW Wind + 33GWh Deferrable Demand
Case 4: Base + 29GW Wind + 33GWh Collocated Storage
Cases 3 & 4 v Case 2: More wind generation is dispatched and the daily dispatch profile of conventional generating units is flatter. Case 3 v Case 4: More wind generation is dispatched in Case 4 BUT the peak system load is lower in Case 3 less congestion on the grid and less utility-owned capacity needed to maintain System Adequacy lower capital costs lower bills for customers.
Environmental Analysis (to be completed)
• Initially, fixed coefficients were used to link generator outputs and emission rates in the SuperOPF, but the emission coefficients for carbon are now being modified to depend on the generator types and operating conditions.
• Time-dependent and location-dependent damage coefficients are being developed to make it feasible to determine how dispatch patterns could be modified to reduce the severity of air pollution (e.g., ozone episodes).
• Using storage and/or deferrable demand may be effective ways to mitigate ozone episodes and increase social benefits by discharging more energy during critical periods when the damage coefficients of the precursor emissions are high (primarily NOx).
51
Potential Benefits
1) Provide a comprehensive analytical framework for evaluating how deferrable demand (electric vehicles and buildings with thermal storage) can affect the operations and costs of an electric delivery system
2) Demonstrate how deferrable demand can: • Flatten the daily dispatch pattern of conventional generators, • Mitigate the variability of wind generation, • Reduce ramping costs and maintain reliability, • Lower costs to customers, • Improve environmental quality (to be completed).
3) The software is open source, and when it is sufficiently robust, it will be added to the programs available with MATPOWER <http://www.pserc.cornell.edu/matpower/>.
52
Publications 1. Lamadrid, Alberto; Wooyoung Jeon, and Tim Mount. The Effect of Stochastic Wind Generation on Ramping Costs and the System Benefits of Storage. Proceedings of the 46th IEEE HICSS Conference, Maui, HI, January 2013. 2. Jia, Liyan; Lang Tong and Tim Mount. Pricing and Competition for Large Scale Charging of Electric Vehicles. Proceedings of the 46th IEEE HICSS Conference, Maui, HI, January 2013. 3. Mount, Tim, Judy Cardell, Lindsay Anderson and Ray Zimmerman, Coupling Wind Generation with Controllable Load and Storage: A Time-Series Application of the SuperOPF, Final Project Report for Project M-22, PSERC Publication 12-29, Nov. 2012. 4. Valentine, K.; W. Temple, and K. M. Zhang. Electric Vehicle Charging and Wind Power Integration: Coupled or Decoupled Electricity Market Resources? Proceedings of the 2012 IEEE PES General Meeting, San Diego, CA, July, 2012. 5. Lamadrid, Alberto. Alternate Mechanisms for Integrating Renewable Sources of Energy into Electricity Markets. Proceedings of the 2012 IEEE PES General Meeting, 22 - 26 July 2012, San Diego, CA, July 22-26, 2012. 6. Lamadrid, Alberto; and Tim Mount. Ancillary Services in Systems with High penetrations of Renewable Energy Sources, the Case of Ramping. Energy Economics, 34:1959–1971, 2012. 7. Mount, Tim; and Alberto Lamadrid. Using Deferrable Demand to Increase Revenue Streams for Wind Generators. Proceedings of the 25th Annual CRRI Western Conference, Monterey CA, June 27-29, 2012. 8. Mount, Tim; Alberto Lamadrid, Wooyoung Jeon, and Hao Lu. Is Deferrable Demand an Effective Alternative to Upgrading Transmission Capacity? Proceedings of the 31st Annual CRRI Eastern Conference, Shawnee, PA, May 16-18, 2012. 9. Palacio, S.; K.; Valentine; M. Wong, K. M. Zhang. System-level price responsive ice storage systems. World Renewable Energy Forum, Denver, Colorado, May 2012. 10. Valentine, K.; W. Temple, and K. M. Zhang. Intelligent electric vehicle charging: Rethinking the valley fill. Journal of Power Sources, 196 (24): 10717-10726, 2011. 11. Mount, Timothy D.; Alberto J. Lamadrid, Surin Maneevitjit, Robert Thomas, and Ray Zimmerman. The Hidden System Costs of Wind Generation in a Deregulated Electricity Market. Journal of Energy Economics, 33 (1), 173-198, 2011. 12. Mount, Timothy D.; Alberto J. Lamadrid, Wooyoung Jeon, and K. M. Zhang. The Potential Benefits for Electricity Customers from Controllable Loads. Proceedings of the 24th Annual CRRI Western Conference, Monterey, CA, June 2011.
53
Task 4: Stochastic Simulation of Power Systems with Integrated
Variable Energy Resources
George Gross Alejandro Dominguez-Garcia
Yannick Degeilh University of Illinois
at Urbana-Champaign
54
THE NEED FOR A NEW SIMULATION APPROACH
Conventional probabilistic simulation cannot
capture the time-varying nature and the inter-
temporal effects that characterize storage and
renewable resources, nor the impacts of
transmission constraints on market outcomes
Moreover, the detailed representation of such
features is analytically intractable 55
THE NEED FOR A NEW SIMULATION APPROACH
We developed a new, comprehensive simulation
approach using stochastic process representation
in a Monte Carlo framework
Our simulation approach is able to represent the
uncertainty in and time–varying nature of the
loads, renewable resource outputs and
conventional unit available capacities, as well as
the time-dependent transmission impacts 56
THRUST OF THE SIMULATION APPROACH
We developed a comprehensive, computationally efficient Monte Carlo simulation approach to emulate the behavior of the power system with integrated storage and renewable energy resources
We model the system load and the resources by discrete-time stochastic processes
We use a storage scheduler to exploit arbitrage opportunities in the storage unit operations
We emulate the transmission-constrained hourly day-ahead markets (DAMs) to determine the power system operations in a competitive environment
57
SIMULATION APPROACH: CONCEPTUAL STRUCTURE
“dri
ver”
stoc
hast
ic p
roce
sses
renewable power outputs
conventional generator available capacities
loads
storage
schedule
market
clearing
procedure
(DCOPF)
congestion rents
CO 2
emissions
storage
operations
. . .
“out
com
e” st
ocha
stic
pro
cess
es
LMPs
58
THRUST OF THE APPROACH
We collect sample paths of the market outcome
stochastic processes to evaluate the expected
system variable effects
Metrics we evaluate include:
nodal electricity prices (LMPs)
generation by resource and revenues
congestion rents
CO 2 emissions
LOLP and EUE system reliability indices 59
KEY CONTRIBUTIONS Development of an effective simulation approach
able to address power industry challenges
Salient features include:
quantification of the power system expected
variable effects – economics, reliability and
environmental impacts – in each sub-period
computationally tractable for practical
systems 60
KEY CONTRIBUTIONS detailed stochastic models of the time–varying
resources and loads allow the representation of
spatial and temporal correlations
storage scheduler for optimized storage
operation to exploit arbitrage opportunities
representation of the transmission–
constrained market outcomes
flexibility in the representation of the market
environment/policies in effect 61
APPLICATION AREAS Resource planning studies
year of commissioning of a wind farm or storage plant
siting of a storage unit transmission utilization under deepening
ADRR implementation Production costing issues
impacts of various penetration levels of wind and/or storage resources
impacts of increases in fossil fuel prices 62
APPLICATION AREAS
Transmission utilization issues
impacts of renewable storage integration on
transmission utilization
identification of frequently-congested
transmission lines for use in the construction
of portfolios of financial transmission rights 63
APPLICATION AREAS
Environmental assessments
identification of appropriate generation
resource mix composition to reduce CO2
emissions by x % by a specified point in time
wind and storage resource synergies in terms
of CO2 emission impacts 64
APPLICATION AREAS Reliability analysis
assessment of the effective load carrying
capability of renewable resource additions
requirements
evaluation of the reserves in a power system
with deepening levels of renewable penetration 65
APPLICATION AREAS Investment analysis
assessment of the expected returns of wind
resources investments
risk assessment of investing in deeper
penetration of renewable resources 66
APPLICATION AREAS Policy formulation and analysis
incorporation of a ‘cap and trade’ carbon
market for the US
assessment of the impacts of a policy aimed at
providing financial incentives for the
retrofitting of old generation units 67
APPLICATION AREAS Broad range of questions
sensitivity studies on storage sizing
selection of remuneration schemes for DRRs
scenario analysis of the impacts of future
technology developments
various what if questions 68
CASE STUDY : DEEPENING WIND PENETRATION
The objective of this study is to perform a wind
penetration sensitivity analysis and to quantify
the enhanced ability to harness wind resources
with the addition of a storage energy resource
We evaluate the key metrics for variable effect
assessment, including wholesale purchase
payments, reliability indices and CO 2 emissions 69
THE STUDY TEST SYSTEM: A MODIFIED IEEE 118-BUS SYSTEM
Annual peak load: 8,090.3 MW
Conventional generation resource mix: 9,714 MW
4 wind farms located in the Midwest with total
nameplate capacity in multiples of 680 MW
A storage unit with 400 MW capacity, 5,000 MWh
storage capability and 89 % round-trip efficiency
Unit commitment uses a 15 % reserves margin provi-
ded by conventional units and the storage resources
Wind power is assumed to be offered at 0 $/MWh
70
NODE 80 AVERAGE HOURLY LMPs LM
P in
$/M
Wh
hour
0 MW 680 MW 1,360 MW
2,720 MW
2,040 MW
0 24 48 72 96 120 144 168 0
5
10
15
20
25
30
35
40
45
50 with storage without storage
wind nameplate capacity
71
EXPECTED WHOLESALE PURCHASE PAYMENTS
thou
sand
$
100
110
120
130
140
150
160
170
180
- 5.
1 %
- 9.
1 %
- 17
.9 %
- 25
.2 %
- 31
.1 %
- 15
.4 %
- 24
.4 %
- 31
.2 %
- 36
.9 %
with storage without storage
72
EXPECTED CO 2 EMISSIONS
- 5.
6 %
- 10
.5 %
- 14
.5 %
- 17
.5 %
- 6.
0 %
- 11
.8 %
- 16
.4 %
- 19
.8 %
0.70
0.75
0.80
0.85
0.90
0.95
1.00
1.05
+ 0
.3 %
thou
sand
met
ric
tons
with storage without storage
73
ANNUAL RELIABILITY INDICES
EUE
6.24
0
1.04
2.08
3.12
4.16
5.20
0
2
4
6
8
10
12
x 10-4
- 37
.5 %
- 50
.0 %
- 65
.6 %
- 80
.1 %
- 43
.8 %
- 56
.3 %
- 67
.1 %
- 71
.9 %
- 87
.5 %
- 42
.5 %
- 73
.9 %
- 80
.2 %
- 64
.1 %
- 55
.1 %
- 90
.0 %
- 82
.4 %
- 73
.5 %
- 94
.3 %
14
LOLP
with storage without storage with storage without storage
MWh
74
CONCLUDING REMARKS
Storage and wind resources consistently pair well
together: they reduce wholesale purchase dollars
and improve system reliability; storage seems to
attenuate the “diminishing returns” trend seen
with deeper wind power penetration
Additional studies are needed to evaluate the
impacts of multiple storage units on the power
system variable effects and the impacts of
storage siting and sizing 75
REFERENCES Y. Degeilh and G. Gross, “Stochastic Simulation of
Power Systems with Integrated Intermittent Energy Resources,” under preparation for submission to the IEEE Transactions on Power Systems, 2013.
Y. Degeilh and G. Gross, “Simulation of Energy Storage Plants in Power Systems with Integrated Intermittent Energy Resources,” under preparation for submission to the IEEE Transactions on Power Systems, 2013.
Y. Degeilh, “Stochastic Simulation of Power Systems with Variable Energy Resources,” thesis to be submitted to the Department of Electrical and Computer Engineering, University of Illinois, Urbana, IL to fulfill the Ph. D. degree requirements, 2013.
76
Contact Information
Shmuel Oren ([email protected]) Duncan Callaway ([email protected])
George Gross ([email protected]) Tim Mount ([email protected])
77